| 一种基于类融合向量的支持向量机及其在语音识别系统中的应用 |
| |
| 引用本文： | 蔡铁,朱杰.一种基于类融合向量的支持向量机及其在语音识别系统中的应用[J].电路与系统学报,2005,10(6):106-110. |
| |
| 作者姓名： | 蔡铁 朱杰 |
| |
| 作者单位： | 上海交通大学,电子工程系,上海,200030 |
| |
| 基金项目： | 上海市科委资助项目，美国贝尔实验室基金 |
| |
| 摘 要： | 支持向量机可以通过产生的支持向量来概括数据集合中的信息,其分类函数仅依赖于一小部分训练样本,即支持向量,这使得它对噪声数据非常敏感.本文采用数据融合的方法加以改进,提出了一种新的基于类融合向量的支持向量机,降低了对噪声数据和较大偏差值的敏感性,提高了算法的容噪性能,并成功地应用于语音识别系统中,取得了较好的效果.
|
| 关 键 词： | 模式识别 数据融合 支持向量机 分布图法 类融合支持向量机 |
| 文章编号： | 1007-0249（2005）06-0106-05 |
| 收稿时间： | 2003-09-24 |
| 修稿时间： | 2004-02-19 |
An improved support vector machine based on class-fusion vectors and application in speech recognition system |
| |
| CAI Tie,ZHU Jie.An improved support vector machine based on class-fusion vectors and application in speech recognition system[J].Journal of Circuits and Systems,2005,10(6):106-110. |
| |
| Authors: | CAI Tie ZHU Jie |
| |
| Abstract: | Support vector machine can use the support vectors to represent all the information about classification in the training set. Its final classification function only depends on a small part of the training samples (the support vectors), which may make it too sensitive to noise and outliers. This paper improves the algorithm through the data fusion method and proposes a new support vector machine based on class-fusion. The theoretical analysis and experimental results show that this new algorithm is more robust, especially when the data distribution of sample set is complicated. |
| |
| Keywords: | pattern recognition data fusion support vector machine distribution display method class-fusion support vector machine |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
